The Silver Plated Lead Bullet
Organizations, attempting to either cut costs or to streamline their information systems development have been outsourcing. That is, contracting out the organization’s project leaders, systems’
analysts, programmers, and data processing operations. This however, is like “getting rid of centuries accumulated know-how (1). This quote from Strassmann really gets to the core of the problem with outsourcing. Accumulated business wisdom is being tossed out the window.”
Database is policy organization and management. DBMS is the technology based implementation of database. Given that data is executed policy, then data’s specification is the technology-based
specification of the corporation’s long-term memory. How can businesses out-source the very definition of its corporate memory to individuals who have no business-wisdom? That doesn’t make
nonsense, it makes business suicide.
If outsourcing is inevitable then businesses must demand that every data processing task be accomplished such that the delivered system be free from all outsourcing vendor’s proprietary CASE, that
the computer languages be ANSI standard, that the DBMS be ANSI standard SQL, and that all documentation, from initial business policy and procedure down through detailed system documentation be
delivered in a meta data repository that allows full inspection, quality control, configuration management, and outsourcer independent evolution and maintenance.
Stories abound about companies who were so locked into the outsourcer that the business was completely unable to evolve because it had trashed all its professional staff and no longer had any data
processing or systems documentation. Stories also circulate about companies who have outsourced overseas. While their costs were immediately reduced, when it came time to upgrade they couldn’t.
Nobody knew the what nor the how!
Outsourcing firms make their money NOT on problems avoided, but on problems solved. How can you bill for time not expended? Outsourcers only provide the least qualified staff they can get-away
with. That is, minimum business knowledge, minimum education and training, and minimum prior project experience, etc. In addition to assigning the least qualified staff, outsourcing firms generally
do not create the outsourcer-independent critical meta-data that fully frame the business policy and procedure from which databases and systems are built. That’s because the outsourcing firm is
being paid solely on the quantity of systems produced, enhancements built, and/or problems repaired. “Damn the analysis and design, we’re only being paid to code,” one outsourcer was heard to
say!
If outsourcing firms were true engineers, they would be earning their keep on the basis of problems avoided, solutions that have flexibility, and systems that are maintainable for a fraction of
their cost rather than multipliers. But, if an outsourcing firm did the latter rather than the former, they would not be able to keep THEIR costs low. AND, their billing would fall off because they
would not be needed. Clearly such behavior would not be in the interest of the outsourcer’s owners/stockholders!
Benefits from Outsourcing?
What is the benefit to outsourcing? Clearly, one main reason is to reduce the life-time cost of an employee. To wit:
- If average salary of an data processing professional is $60,000 and if G&A; and overhead multiplier is 2.5, then the per year staff cost is $150,000. From 22 to 55 the total cost of the data
processing professional is $4,950,000. If the staffer then draws a retirement of $40,000 and lives 30 more years, the retirement cost is 1,200,000. The total staff cost is thus, $6,150,000. - Alternatively, if a data processing professional is contracted through an outsourcer at $100/hour then the annual cost is $184,000. Over 30 years the total cost is $5,520,000. The possible
savings approximate $600,000 per person. In addition, business’ can trade in the outsourcing staffer every year for better and newer models.
Saving money is however is a subterfuge because with outsourcing, there is no real retention of corporate wisdom through the building of a corporate meta data meta data repository. The value of
this however, is both difficult to quantify and at times, completely unvalued because there is no acceptance that the residual products of a data processing professional is an asset. Sadly, the
only identified and valued asset is the system/program produced. Businesses have no understanding of value derived from leveraging long-term corporate knowledge into building more and better
systems/programs at both an increased quality and faster rate.
To make the outsourcer’s work valuable, a meta-data meta data repository must be fully defined and put into place as a minimally acceptable work environment for the outsourcer’s effort. The
United States Army in the mid Eighties contracted for the development of a world wide system to collect maintenance data on targeted sets of material to predict reliability, availability, and
maintainability. Ten systems had to be built in less time than the previous two. The design and build was accomplished on-time and for less funds than were predicted simply because of the meta data
repository which stored the reusable meta data.
The Payoff
The U.S. Army had systems that did not cost 10 * 100%, or $5 million. Rather, they had systems that cost 64% less, or $1.8 million. A savings of $3.2 million. The reasons were simple:
- A well defined database oriented methodology that clearly set down procedures, specified products, and quality reviews.
- A meta data repository that enabled the continued use and reuse of meta data to accelerate development and maintenance.
The U.S. Army had outsourced the entire data collection, analysis, and reporting effort to an Iowa defense contractor specializing in logistics analysis and planning. The contractor, in turn,
contracted out all data processing system analysis, design, implementation to other contractors. The contractor operated the logistics system itself and kept the meta data repository “inside”
their domain. This enabled the contractor to be independent of its own subcontractor vagaries by exerting positive control over all planning, requirements, quality control, all phases of testing,
and complete design through user documentation.
Within a year the contractor no longer needed its subcontractors because all the systems’ meta data was designed and implemented through CASE, repositories, and code generators that the contractor
owned. The Army, on the other hand was not as good a buyer as the contractor because when the contract ended, the Army did not own the meta data repository. The Army had only contracted for the
delivery of the software systems and all pertinent documentation. The contractor complied by delivering a small “dump truck” of documents representing the source code and executable forms of
software. Not delivered were the two meta data repository-based factories: analysis and design, and software implementation and maintenance. While the source and executable software and
documentation were the “golden eggs,” the two meta data repository-based factories were the “goose.” It was because of these two factories that the contractor was able to deliver 10 systems in
one year rather than the two systems in three years.
The Army did not comprehend the value and the critical role of the meta data repository. Without the meta data repositories the new contractor could not create new systems nor evolve existing
systems. Slowly, but surely the entire systems environment crumbled and was discontinued. Without the CASE, meta data repository, and code generator environment there were no more golden eggs; the
goose was gone. Scourge or salvation? It depends significantly on who defines, owns, and controls the assets. That is, the CASE, meta data repository, and code generator environment. To one,
outsourcing was salvation. To the Army it was a scourge.
(1) – This quote was taken from an article by Paul Strassman (former CIO of major U.S. corporations and the U.S. DoD) entitled Outsourcing Game for Losers,
August 21, 1995 in Computer World.
